Why People Are Freaking Out About Clawdbot
Transcript and Summary Unavailable
The transcript for this video is currently unavailable. Likely due to restrictions set by the content owner.
Related videos
AI News: These Google Updates Are Dividing People
Matt Wolfe
38.6k views
Heinz CEO DIRE WARNING: People Running Out Of Money
Breaking Points
258.0k views
AI Is Frying Your Brain
Matt Wolfe
31.7k views
Breaking: Google Upgraded Nano Banana Again!
Matt Wolfe
44.7k views
Are People Actually Getting Dumber? Yes, And Here's Why.
Matt Walsh
25.3k views
The Truth About White People No One Wants To Hear
Matt Walsh
85.9k views
The Clawdbot Story Just Took a WILD Turn
Matt Wolfe
95.7k views
Clawdbot is about to BREAK EVEREYTHING
Wes Roth
12.9k views
Are Black Women The Most Privileged Group Of People Out There?
Matt Walsh
10.2k views
Why So Many People Were Wrong About Altcoins
Benjamin Cowen
52.3k views
Top Comments (10)
Since it can install stuff on its own based on just a message, prompt injection (via website it ends up on via search or by following links) could have it install arbitrary script to exfiltrate your entire history/memory (like API keys, personal information). This is a security nightmare.
Wish I had something for it to do, my life is so simple that I don’t need it.
Skynet comes Undercover before GTA 6 is crazy
Little tip for anyone trying this out. I have this running on my home lab with guard rails, to keep costs down you don't need to use a API connection. For testing I have been using my Chatgpt subscription as this is mentioned to be possible in the documentation. This means your not getting hit for each API call as it's like your using chatgpt yourself. Happy to explain this to anyone that wants to know.
Please read these resources about security concerns and what to do about them if you're going to move forward with your own Clawdbot install: - https://docs.clawd.bot/gateway/security - https://www.perplexity.ai/search/what-are-the-security-risks-of-a3Wkk0bVSMesrZXVp72Vqw#1 - https://twitter.com/themichaelcrist/status/2015771278590755069 - https://twitter.com/ItakGol/status/2015848329351958767 NOTE: There have been some talk floating around that Anthropic is stopping people from using their Claude Max plans. They want you to use the unmetered API instead. Here are the resources mentioned in the video (Give them a like): EC2 Setup Instructions: https://x.com/damianplayer/status/2015105669620269373?s=12 EC2 Setup Instructions: https://x.com/techfrenaj/status/2014934471095812547?s=12 Local Setup Instructions: https://x.com/AlexFinn/status/2015182480064893118 Local Setup Instructions: https://x.com/nickspisak_/status/2015144613594415164?s=12 Tips: https://x.com/altryne/status/2015222870591553559?s=12 Resource Roundup: https://x.com/scobleizer/status/2015346738517942628?s=12 What They're Not Saying: https://x.com/heyshrutimishra/status/2015327280911073789?s=12 Security is Scary: https://x.com/rahulsood/status/2015397582105969106?s=12
Mad respect for that disclaimer. Those warnings are something you see very few AI bros actually do.
Interested in seeing more professional, specific, use-cases. Not overly excited by any that I've seen yet.
Wow, Maybe its too cool to be true so ill hold off a couple weeks before installing it and watch the threads for any security problems. Hopefully trustable people like you could give us an update if ya find anything a little funny going on. Im sure everyone would appreciate that. Thanks Matt
The only addition I would make to this lovely tutorial is when you create an AWS instance, always create a private key ".ppk" and then use tools like Putty to SSH into your instance and from there run all your Linux commands or do whatever you are supposed to do with your instance.
Ok. Looks like a got a project when I get home tonight. Thanks Bud. I like it.
Unlock the Data Inside
Turn Videos into Knowledge
- Get FREE 10/day: transcripts, summaries, chats
- Chat with videos, export text & PDF
- $1 free API credit for RAG, chatbots & research
Free forever plan • All features unlocked
Top Comments (10)
Since it can install stuff on its own based on just a message, prompt injection (via website it ends up on via search or by following links) could have it install arbitrary script to exfiltrate your entire history/memory (like API keys, personal information). This is a security nightmare.
Wish I had something for it to do, my life is so simple that I don’t need it.
Skynet comes Undercover before GTA 6 is crazy
Little tip for anyone trying this out. I have this running on my home lab with guard rails, to keep costs down you don't need to use a API connection. For testing I have been using my Chatgpt subscription as this is mentioned to be possible in the documentation. This means your not getting hit for each API call as it's like your using chatgpt yourself. Happy to explain this to anyone that wants to know.
Please read these resources about security concerns and what to do about them if you're going to move forward with your own Clawdbot install: - https://docs.clawd.bot/gateway/security - https://www.perplexity.ai/search/what-are-the-security-risks-of-a3Wkk0bVSMesrZXVp72Vqw#1 - https://twitter.com/themichaelcrist/status/2015771278590755069 - https://twitter.com/ItakGol/status/2015848329351958767 NOTE: There have been some talk floating around that Anthropic is stopping people from using their Claude Max plans. They want you to use the unmetered API instead. Here are the resources mentioned in the video (Give them a like): EC2 Setup Instructions: https://x.com/damianplayer/status/2015105669620269373?s=12 EC2 Setup Instructions: https://x.com/techfrenaj/status/2014934471095812547?s=12 Local Setup Instructions: https://x.com/AlexFinn/status/2015182480064893118 Local Setup Instructions: https://x.com/nickspisak_/status/2015144613594415164?s=12 Tips: https://x.com/altryne/status/2015222870591553559?s=12 Resource Roundup: https://x.com/scobleizer/status/2015346738517942628?s=12 What They're Not Saying: https://x.com/heyshrutimishra/status/2015327280911073789?s=12 Security is Scary: https://x.com/rahulsood/status/2015397582105969106?s=12
Mad respect for that disclaimer. Those warnings are something you see very few AI bros actually do.
Interested in seeing more professional, specific, use-cases. Not overly excited by any that I've seen yet.
Wow, Maybe its too cool to be true so ill hold off a couple weeks before installing it and watch the threads for any security problems. Hopefully trustable people like you could give us an update if ya find anything a little funny going on. Im sure everyone would appreciate that. Thanks Matt
The only addition I would make to this lovely tutorial is when you create an AWS instance, always create a private key ".ppk" and then use tools like Putty to SSH into your instance and from there run all your Linux commands or do whatever you are supposed to do with your instance.
Ok. Looks like a got a project when I get home tonight. Thanks Bud. I like it.